Automating the Art of Jalebi Manufacturing: A Machine Learning-Assisted Inverse Kinematics Approach for Precision Path Planning and Scalable Production of a Traditional Indian Sweet

Jalebi is a popular deep-fried Indian dessert made from fermented batter, typically consisting of refined flour and yogurt. The automation of jalebi making is essential to improve efficiency, consistency, and hygiene in large-scale production. Traditional manual preparation lacks hygiene, requires intensive labor, consumes time, and results in inconsistencies in shape, size, and texture. An automation in jalebi-making ensures uniform batter dispensing, precise frying, and controlled sugar syrup absorption, reducing human effort and production costs. Additionally, automation enhances food safety by minimizing direct human contact, making it an ideal solution for commercial kitchens, sweet shops, and industrial-scale food production. The automation of Jalebi production presents a significant challenge due to intricate motion patterns traditionally performed by skilled artisans. Conventional inverse kinematics approaches struggle to replicate these patterns due to the absence of a mechanistic motion model. This study proposes a novel machine learning-based path planning method for the automated production of Jalebi, ensuring consistent shape, quality, and texture while preserving traditional craftsmanship. Motion data was collected from skilled artisans to capture path patterns, size variations, and production speeds. A parametric equation was developed to represent the Jalebi formation process, enabling the formulation of an inverse kinematics solution for an RUU-parallel robot configuration. To the best of our knowledge, this is the first attempt at automating Jalebi making using an RUU-parallel manipulator, a low cost solution. A supervised learning model was trained on these motion patterns, achieving a mean squared error of 1.005 × 10−12 after 1000 iterations and an R-squared value of 1. The trained network was validated which shows the mean square error of 0.0988 and an optimized R-squared value 0.8742. The approach was validated through simulations, demonstrating accuracy, consistency, and adaptability across different Jalebi designs and production scales. This framework enables scalable, precise, and efficient Jalebi manufacturing, with potential applications in automating other intricate food production processes, such as Chakari and Imarti, thereby bridging cultural heritage with modern automation technologies. In addition, the conceptualized Automatic Jalebi Making Machine not only optimizes the automation process but also provides a cost-effective and scalable solution for similar food preparation industries.